Recombinant Expression of a Ready-to-Use EGF Variant Equipped With a Single Conjugation Site for Click-Chemistry

Abstract

The epidermal growth factor (EGF) receptor is commonly targeted in cancer therapy because it is overexpressed in many malignant cells. However, a general problem is to couple the targeting moieties and the drug molecules in a way that results in a homogeneous product. Here, we overcome this issue by engineering a variant of EGF with a single conjugation site for coupling virtually any payload. The recombinant EGF variant K-EGF^RR was expressed in E. coli Rosetta with a 4-6 mg/L yield. To confirm the accessibility of the introduced functional group, the ligand was equipped with a sulfo-cyanine dye with a loading of 0.65 dye per ligand, which enables tracking in vitro. The kinetics and affinity of ligand-receptor interaction were evaluated by enzyme-linked immunosorbent assay and surface plasmon resonance. The affinity of K-EGF^RR was slightly higher when compared to the wild-type EGF (K _D: 5.9 vs. 7.3 nM). Moreover, the ligand-receptor interaction and uptake in a cellular context were evaluated by flow cytometry and quantitative high-content imaging. Importantly, by attaching heterobifunctional polyethylene glycol linkers, we allowed orthogonal click-conjugation of the ligand to any payload of choice, making K-EGF^RR an ideal candidate for targeted drug delivery.

Keywords: click‐chemistry; epidermal growth factor; ligand variant; single conjugation site; targeted therapy.

FIGURE 1

Design of the variant K‐EGF^RR with additional 6 × His‐Tag and its site‐specific modifications. R32 and R52 indicate the lysines, which were replaced by arginine (R) at positions 32 and 52. The artificially inserted lysine serves as a single conjugation site and is indicated in blue, whereas the His‐tag is depicted in yellow. The EGF variant was equipped with dye and non‐cleavable or cleavable PEG linkers by conjugating them at the single conjugation site. The schemes are not to scale.

FIGURE 2

Successful expression leading to the pure K‐EGF^RR. (A) Expression in different E. coli strains. (B) Elution of K‐EGF^RR expressed in Rosetta(DE3) after NiNTA affinity chromatography and (C) conjugation of Cy3‐dyes to K‐EGF^RR to confirm the accessibility of the functional groups.

FIGURE 3

Successful conjugation of orthogonal linkers to K‐EGF^RR. Conjugation of (A) non‐cleavable and (B) cleavable DBCO‐PEG‐NHS linker at K‐EGF^RR with different equivalents of the linker. Afterward, the K‐EGF^RR‐linker samples were incubated with Cy5‐N₃ to assess whether the linker was successfully conjugated. The images were obtained by using the Cy5‐channel of a VersaDoc device.

FIGURE 4

Higher EGFR binding affinity of K‐EGF^RR than commercially available wild‐type EGF determined by ELISA. Biotinylated EGFR was immobilized on streptavidin‐coated well plates and treated with K‐EGF^RR/EGF. Absorbance was measured at 450 nm. The mean value ± SEM is shown (N = 3, n = 3; N: number of biological replicates, n: number of technical replicates per biological replicate).

FIGURE 5

Higher ligand–receptor affinity in the case of K‐EGF^RR compared to commercially available wild‐type EGF determined by SPR. Biotinylated EGFR was immobilized on a streptavidin‐coated SPR chip and treated with different amounts of (A) K‐EGF^RR or (B) commercially available EGF. (C, D) The equilibrium dissociation constant K _D was determined by using the steady‐state affinity fit. The mean value ± SEM is shown. (N = 1, n = 2).

FIGURE 6

Receptor‐mediated uptake of the Cy5‐labeled EGF variant determined by flow cytometry. K‐EGF^RR was evaluated in cell lines A2058 (EGFR–), MDA‐MB 468 (EGFR+) and A431 (EGFR++) (0.25 Cy5/EGF). (A) Concentration and (B) time‐dependent uptake of Cy5‐labeled K‐EGF^RR incubated for 24 h. (C) Competitive inhibition of the internalization of Cy5‐labeled K‐EGF^RR by unlabeled K‐EGF^RR in varying concentrations. The mean value ± SEM is shown (N = 3, n = 2).

FIGURE 7

Receptor‐mediated endocytosis of EGF wild‐type and variant. (A) High‐throughput imaging of HeLa Kyoto cells either transfected with mock or siRNA against EGFR. Cells were treated with commercial EGF‐labeled with Alexa‐647 [EGF (wt), 9 nM], (sulfo‐Cy5‐NHS)0.25‐K‐EGF^RR [K‐EGF^RR, 9 nM] or left untreated [DMEM] for 0, 10, 30, 60, 120 and 300 min (green). Cells were stained with a mix of DAPI and CellMask (grey). (B) Plot of the percentage of EGF‐positive cells for EGF (wt) and K‐EGF^RR in relation to overall cells for different time points and on cells treated with EGFR siRNA and mock. (C) Plot of the mean number of overall cells treated with different controls. The mean value ± SEM is shown (N = 4, n = 4).